文档介绍：Proceedings of the American Control Conference
Anchorage, AK May 8-10,2002
Behavior-Based Learning Fuzzy Rules for Mobile Robots
Siripun Thongchai
Department of Teacher Training in Electrical Engineering
King Mongkut’s Institute of Technology North Bangkok, Bangkok, Thailand 10800
{srpQkmitnb . ac .th}
Abstract In this work, a learning system for a mobile robot is
designed based on fuzzy control technique. Laser and
This paper describes a method to build a pass readings are used as the fuzzy inputs. The
system for mobile robots based on fuzzy control tech- learning system generates a set of fuzzy rules for build-
nique. Range sensors are analyzed and used. The learn- ing behaviors of mobile robots. Two examples, avoid-
ing system generates a set of fuzzy rules based on the obstacle and detect-landmark behaviors are illustrated.
robot’s behaviors. Two examples illustrate the learning Each set of fuzzy rules is produced based on the sensory
capabilities: 1) learn-to-avoid-obstacle and 2) learn-to- information and motor action which are related to the
detect-landmark behaviors. Reinforcement learning is robot’s behaviors. The results are used for mobile robot
applied to correct the fuzzy rules. Fuzzy controllers navigation and tested using an ATRV-Jr mobile robot
are used to control the robot using a set of fuzzy rules.
131.